The day's AI news, briefed
The Daily Prompt
Our AI editor reads each day's AI and robotics stories, pulls out why they matter, and links every original source — so you can keep up in minutes.
Join over 250,000 subscribers getting the latest AI news
Latest AI & Robotics News Today
Google’s more affordable AI Plus plan rolls out to all markets, including the U.S.
The Google AI Plus plan's global expansion aims to provide more accessible AI tools, potentially increasing AI technology adoption among casual users and emerging markets.
- Google's AI Plus plan is now available in all markets, including the U.S., for $7.99 per month.
- The expansion brings the plan to 35 new countries following earlier launches that began last September.
- The plan includes access to tools like Gemini 3 Pro, Flow’s AI filmmaking tools, and 200GB of storage.
Google's AI Plus plan is a subscription offering designed to provide affordable access to AI tools. Launched initially in emerging markets, it aims to attract users who may not afford or need the more expensive AI Pro plan.
Read the original ↗
AI Overviews gets upgraded to Gemini 3 with a dash of AI Mode
The upgrade to Gemini 3 improves the accuracy of AI Overviews, impacting how users interact with Google's search services and highlighting the evolution of AI in search technology.
- Google's AI Overviews is being upgraded to Gemini 3 models, offering more accurate responses.
- Gemini 3 improves upon the previous Gemini 2.5 model in knowledge-based benchmarks, achieving better results.
- AI Overviews can now transition to AI Mode for a conversational search experience, enhancing user interaction.
Google's Gemini models are generative AI systems designed to provide answers based on user queries. The previous version, Gemini 2.5, was useful but less accurate compared to Gemini 3. Future developments to watch include further integration of AI Mode and its impact on traditional search results.
Read the original ↗
Google now lets users jump from AI Overviews into AI Mode conversations
This update enhances how users interact with Google Search, shifting it towards a more conversational model that can handle complex queries more effectively.
- Google now allows users to ask follow-up questions from AI Overviews, facilitating a conversation with its AI Mode for in-depth inquiries.
- Gemini 3 has been announced as the default model for AI Overviews globally, aimed at providing superior responses in search results.
- These enhancements are part of Google's strategy to make Search more interactive and less static, aiming for a seamless user experience.
Google is a major technology company known for its search engine and various digital services. The introduction of AI Overviews and AI Mode represents a shift towards integrating AI into everyday search functionality. Future updates may continue to enhance personal AI interactions across Google's ecosystem.
Read the original ↗
Risotto raises $10M seed to use AI to make ticketing systems easier to use
The funding will enable Risotto to enhance ticketing systems, which could streamline help desk operations and improve customer support efficiency across various industries.
- Risotto has raised $10 million in a seed round led by Bonfire Ventures, with participation from several other investors.
- The startup aims to automate help desk ticket resolution using AI, targeting inefficiencies in current systems.
- Risotto successfully automated 60% of support tickets for the payroll company Gusto, indicating significant operational improvements.
Risotto targets the help desk automation industry, which is currently dominated by larger companies like Zendesk and ServiceNow. The company is developing AI-driven solutions to streamline ticket management, with a focus on integrating with existing systems. Future developments may include leveraging large language models for more effective ticket resolution.
Read the original ↗
Airtable gets into the AI agent game with Superagent
The launch of Superagent highlights Airtable's strategic pivot towards AI, demonstrating the growing emphasis on AI-driven solutions in software development despite significant challenges in company valuation.
- Airtable launches Superagent, its first stand-alone product in 13 years, signifying the company's new focus on AI technology.
- The company's valuation has dropped from $11.7 billion in 2021 to about $4 billion, yet it remains financially stable with significant funds available.
- Superagent employs a multi-agent coordination system that allows it to deploy specialized agents in parallel, enhancing response efficiency for user queries.
- Superagent will operate independently, led by former members of the AI agent startup DeepSky, which Airtable acquired.
- Pricing is likely to start at $20 per month per user, following trends seen in emerging AI products.
Airtable is a no-code platform enabling users to build custom applications easily. The company has transformed by introducing AI capabilities, evidenced by its recruitment of a new CTO from OpenAI and the launch of Superagent as a pivotal tool in its strategy.
Read the original ↗
Phoebe Gates and Sophia Kianni’s Phia raises $35M to ‘make shopping fun again’
The success of Phia, with its innovative AI shopping agent, could reshape online commerce by making shopping more personalized and potentially more efficient for consumers.
- Phia has raised $35 million in funding, led by Notable Capital, to enhance its AI-driven shopping platform.
- Since its launch 10 months ago, Phia has experienced 11x revenue growth and onboarded 6,200 retail partners.
- The app focuses on helping users find resale or secondhand product alternatives to save money, promoting sustainable shopping practices.
Phia is a startup focusing on creating an AI shopping agent that personalizes the shopping experience. Launched 10 months ago, it has quickly grown its user base and revenue. The company aims to recruit top tech talent to further develop its platform and improve user experience.
Read the original ↗
“Sharktober” Isn’t a Myth: Science Explains Why Shark Bites Spike Every October in Hawaii
Understanding the biological reasons behind increased shark bites in October can improve ocean safety and inform public awareness without inciting unnecessary fear.
- A peer-reviewed study found that shark bites in Hawaii spike every October due to tiger shark biology, not increased human activity.
- Of the 165 unprovoked shark bites recorded in Hawaii between 1995 and 2024, 32 occurred in October, marking a significant seasonal increase.
- Tiger sharks accounted for at least 63% of all bites in October, with larger sharks observed near populated coastal areas during this time.
Tiger sharks are the dominant species involved in shark bites in Hawaii. Previous assumptions linked increased shark bites to human activity or temperature changes, which have now been rejected by this study. Further research is needed to explore the reproductive behaviors of tiger sharks and their impact on shark-human interactions.
Read the original ↗
Previously Unknown Roman Villa Discovered Beneath a Welsh Park is Helping Unearth Britain’s Hidden History
This discovery adds significant depth to our understanding of Roman influence in Wales and highlights the potential for further archaeological insights into Britain's past.
- A Roman villa complex has been found three feet below a park in Margam, Wales, using high-resolution magnetometry and ground radar.
- The site was mapped in 3D by the ArcheoMargam team, which includes Swansea University and Neath Port Talbot Council representatives.
- The villa is unusual as it sits within a defensible enclosure, possibly indicating a connection to earlier Iron Age settlements or later Roman defensive needs.
- Additional findings suggest the presence of a possible agricultural storage facility or post-Roman meeting hall nearby, as well as hints of a bathhouse.
The site represents a significant archaeological find, enhancing knowledge about Roman-era life in South Wales. Previous research in the area has typically focused on other historical periods, making this discovery particularly noteworthy.
Read the original ↗
AI Can Match Average Human Creativity—But We Still Hold the Edge Where It Matters Most, New Study Finds
This study provides empirical evidence that while AI has made significant strides in creative tasks, it still trails behind the most creative humans, highlighting the role of human creativity in innovation.
- A new study involving 100,000 participants shows that AI can match average human creativity but lags behind the best human creatives.
- The study used the Divergent Association Task (DAT) to objectively measure creativity in both humans and AI systems, focusing on divergent thinking.
- AI models like GPT-4 performed well against the average human but consistently fell short compared to the top 50% and top 10% of human performers.
Artificial Intelligence (AI) technology has progressed to create systems that can generate creative outputs like poetry and stories. Previous studies have suggested AI might surpass human creativity, but this study demonstrates a significant gap remains in the highest creative capabilities found in humans. Future research may continue to explore AI's potential in creative tasks and its implications for human-AI collaboration.
Read the original ↗
Node-based design tool Flora raises $42M from Redpoint Ventures
Flora's significant funding round highlights the growing importance of user-friendly design tools in the generative AI landscape, which is reshaping creative workflows across industries.
- Flora, a node-based design tool, raised $42 million in Series A funding led by Redpoint Ventures.
- The platform allows users to create media assets through prompts, using image, text, and video with iterative workflows.
- Flora plans to use the funds to enhance enterprise sales and creative controls, as well as increase marketing efforts.
Flora is a design tool utilized by prominent companies to streamline the creative process using generative AI models. Its previous versions have aimed to simplify the complexities of node-based design, making it accessible to both professionals and novices. Future developments include adding more features for traditional editing and expanding the user base.
Read the original ↗
Self-Refining Video Sampling
This research offers a novel approach to enhancing the realism of video generation, which is essential for applications in entertainment, gaming, and simulation technologies.
- Self-refining video sampling improves motion coherence and physics alignment in video generation.
- The method utilizes a pre-trained video generator as a denoising autoencoder, allowing iterative refinement without new training or external verification.
- This approach incorporates an uncertainty-aware strategy to selectively refine certain regions, minimizing artifacts.
Video generators are algorithms that create video content, but they often struggle with realistic motion and dynamics. Traditional methods require heavy computations and augmented data for training. This new self-refining technique aims to streamline video creation by enhancing existing models.
Read the original ↗
Southwest Airlines Says Bye to Open Seating—and Hello to Boarding Complexity
The shift from open seating to assigned seating by Southwest Airlines highlights the challenge of balancing operational efficiency with profitability in airline boarding practices.
- Southwest Airlines has officially implemented an assigned seating policy, ending its previous open seating approach after fifty years.
- The new boarding method uses the Window-Middle-Aisle, or WILMA, technique, where passengers board by group according to their seat location.
- The policy introduces multiple tiers for boarding preference, prioritizing passengers who pay for extra services, which may complicate the boarding process.
- Southwest expects the introduction of extra legroom seats and other perks to generate an additional $1.5 billion in revenue annually by next year.
Southwest Airlines built its reputation on an innovative open seating policy that favored early check-in passengers. The move to assigned seating aligns the airline more closely with industry standards and aims to increase profits amid a competitive market.
Read the original ↗
Revealed: Leaked Chats Expose the Daily Life of a Scam Compound’s Enslaved Workforce
The exposure of these leaked chats reveals the harsh and manipulative conditions faced by forced laborers in scam operations, highlighting serious human rights abuses and the need for action against such practices.
- A whistleblower provided 4,200 pages of internal communications from a scam compound in Laos, detailing the exploitation of workers forced into committing fraud.
- Workers were held in debt bondage, facing fines and severe punishment for failing to meet scam quotas, often leading to insufficient wages and food deprivation.
- The leaked materials reveal a toxic workplace culture that combines corporate language with threats of violence to manipulate and control the enslaved workforce.
The operation implicated in these leaks is part of a broader trend of 'pig butchering' scams, where victims are lured into fraudulent schemes under the guise of legitimate employment. These types of operations have proliferated in Southeast Asia and have been linked to widespread human trafficking issues. Observers are monitoring the situation to see how this awareness might lead to improved protections for victims and enforcement actions against such scams.
Read the original ↗
How to Use an AI Art Generator to Create Stunning Images (Easy & Fast)
The development of AI art generators like a1.art enhances creative productivity, allowing users to quickly create images while also raising important legal and technical considerations for creators.
- AI art generators enable users to create images rapidly, significantly boosting workflow efficiency for designers and creators.
- a1.art stands out for its user-friendly interface, requiring no prior experience in prompt engineering, making it accessible for beginners.
- The platform offers access to over 28,000 distinct image styles and allows generation of culturally specific designs, such as sarees.
- Legal clarity around using AI-generated art is evolving, urging users to be aware of licensing and usage rights on various platforms.
AI art generators use algorithms to transform ideas into visuals, catering to both personal and commercial projects. As these tools gain popularity, understanding their functionalities and regulations is crucial for users. Ongoing legal discussions focus on copyright and the originality of AI-generated content.
Read the original ↗
‘Among the worst we’ve seen’: report slams xAI’s Grok over child safety failures
The report highlights significant safety shortcomings in xAI's Grok chatbot, raising concerns about child protection in AI technologies and prompting legislative scrutiny.
- xAI's Grok chatbot lacks adequate user identification for minors and frequently generates inappropriate content, making it unsafe for children.
- Common Sense Media's assessment indicates Grok's failures include ineffective safety measures that allow minors access to harmful material, despite the introduction of 'Kids Mode'.
- xAI restricted Grok's image generation features to paying subscribers after user and regulatory backlash, yet many continued to access it for free.
- The nonprofit's testing revealed that Grok outputs content that does not prioritize child safety, including harmful advice and sexualized interactions.
- Legislators like Senator Steve Padilla are advocating for stronger regulations on AI chatbots to enhance protections for minors.
xAI is a company focused on artificial intelligence technologies, with Grok being an AI chatbot released to engage users through text and image generation. The report from Common Sense Media exposes severe risks associated with Grok's functionality, mirroring ongoing concerns about AI's impact on youth safety. Legislative efforts are evolving to establish stricter regulations for AI chatbots, particularly concerning user safety for minors.
Read the original ↗
Qualcomm backs SpotDraft to scale on-device contract AI with valuation doubling toward $400M
The investment by Qualcomm affirms the growing importance of on-device AI for legal workflows, addressing concerns about data privacy and security in sensitive contract management.
- SpotDraft raised $8 million from Qualcomm Ventures to enhance its on-device AI technology for contract review.
- The latest funding round values SpotDraft at approximately $380 million, nearly doubling its previous valuation of $190 million.
- SpotDraft has seen contract processing volumes increase by 173% year-over-year, now handling over 1 million contracts annually.
SpotDraft, launched in 2017, focuses on contract lifecycle management with a privacy-first approach. The investment from Qualcomm aims to support the startup's expansion in regulated sectors where data security is critical.
Read the original ↗
OpenAI spills technical details about how its AI coding agent works
OpenAI's detailed technical disclosure about Codex's operations aids developers in understanding how AI coding agents function and addresses ongoing challenges in AI-assisted programming.
- OpenAI engineer Michael Bolin published a technical breakdown of how the Codex coding agent operates, including its design philosophy and internal mechanisms.
- Codex AI coding agents perform coding tasks with human supervision, addressing a range of tasks from writing code to fixing bugs.
- Bolin's post highlights issues such as the inefficiency of prompt growth and performance challenges due to cache misses and bugs.
- OpenAI's Codex CLI client is open-sourced on GitHub, offering developers transparency that is not available for ChatGPT.
- The 'agent loop' is described as a continuous cycle of user input and model response that drives the coding tasks.
OpenAI’s Codex is an AI system designed for software development, allowing users to write code through a command-line interface. This technical disclosure comes as AI coding tools grow in popularity and usefulness, amid ongoing debate about their reliability. Future posts are expected to cover additional technical specifics about Codex's architecture and performance features.
Read the original ↗
Microsoft Introduces Maia 200, a High-Performance Chip for AI Inference
The introduction of the Maia 200 chip signals a significant advancement in AI infrastructure, enabling faster and more efficient processing for increasingly complex AI models, which impacts the entire AI ecosystem.
- Microsoft has announced the Maia 200, a custom chip designed to improve the efficiency of AI inference.
- The Maia 200 delivers over 10 petaflops of performance at 4-bit precision and 5 petaflops at 8-bit precision.
- This new chip includes more than 100 billion transistors, allowing it to handle heavier AI workloads with less energy consumption.
- Maia 200 aims to lower costs and increase stability across AI systems, marking it as part of Microsoft's long-term AI strategy.
Microsoft's Maia 200 is part of a trend where major tech companies are developing custom AI chips to reduce dependence on Nvidia. The Maia 200 follows the Maia 100, launched in 2023, and aims to enhance AI inference by offering significant performance improvements. The chip is now available for internal workloads and is being opened to external developers and researchers for broader use.
Read the original ↗
YouTubers sue Snap for alleged copyright infringement in training its AI models
The lawsuit reflects ongoing tensions between content creators and companies using AI, particularly over copyright concerns in the evolving landscape of digital media.
- A group of YouTubers have sued Snap for alleged copyright infringement in training its AI models using their video content.
- The plaintiffs claim Snap used a dataset designed for academic and research purposes to develop its AI features like 'Imagine Lens'.
- This lawsuit adds Snap to a list of defendants that includes Nvidia, Meta, and ByteDance, with ongoing disputes regarding AI training practices.
- The case seeks statutory damages and a permanent injunction to prevent further copyright violations by Snap.
Snap is being accused by content creators of misusing copyrighted video material to enhance its AI functionalities. This lawsuit comes amid broader litigation involving AI firms and content creators regarding copyright infringement, signaling a critical legal battle in the tech industry.
Read the original ↗
“Wildly irresponsible”: DOT’s use of AI to draft safety rules sparks concerns
The use of AI by the Department of Transportation for drafting safety rules raises serious concerns about the potential for errors that could lead to severe consequences in transportation safety.
- The US Department of Transportation plans to use AI, specifically Google Gemini, to draft rules impacting transportation safety.
- DOT staffers express skepticism about the AI's ability to handle the complexities of regulatory drafting, emphasizing the need for human expertise.
- Errors in drafted regulations could result in injuries or fatalities, as the rules affect critical areas such as air travel and pipeline safety.
The Department of Transportation's initiative to utilize AI for drafting rules could set a precedent for government agencies. Concerns stem from previous instances where AI has produced inaccurate information, particularly in legal contexts. Stakeholders are watching closely as this plan unfolds amid broader discussions on AI's role in government.
Read the original ↗